Oscillators Clapp Oscillator VFO Startup
Oscillators are circuits that produce periodic output voltages,
such as sinusoids. They accomplish this feat without any “input”
signal, other than dc power. Our NorCal 40A has three:
1. VFO (an LC oscillator),
2. BFO (a crystal oscillator),
3. Transmitter oscillator (also a crystal oscillator).
You’ve likely had some experience with oscillators, perhaps
with the astable multivibrator using the 555 IC. This RC
oscillator produces a square-wave output voltage that is useful at
low frequencies. Generally used in hobby-type circuits.
The oscillators in the NorCal 40A are called feedback
oscillators. This is a somewhat difficult subject since these
oscillators are intrinsically nonlinear devices.
Feedback oscillators have three basic parts:
1. An amplifier with signal gain G,
2. A linear feedback network with signal loss L,
3. A load of resistance R.
We’ll ignore the effects of R for now.
The amplifier and feedback network are connected as shown in
For the amplifier y = Gx
while for the feedback network x =y/l Þ y =Lx
Here we have two equations for two unknowns. However, these
are not linearly independent equations! If
• G ? L ? x = y = 0. No oscillation is possible.
• G = L ? y and x may not be zero. Hence, oscillation is
possible.
Generally, G and L are complex numbers, so we have two real
equations to satisfy in the equality G = L:
1. G = L – the magnitudes are equal .
2. ?G = ?L – the phase angles are equal.
The meaning of (11.3) is that the gain of the amplifier
compensates for the loss of the feedback network. The meaning
of (11.4) is that the feedback network compensates for the phase
shift (i.e., time delay) of the amplifier.
In a feedback oscillator, noise in the circuit will be amplified
repeatedly until a single frequency output signal y is produced –
a perfect oscillation. In general can be satisfied for the situations shown in G may decreases at high power levels due to amplifier overloading:
In the NorCal 40A this decreasing G occurs because of gain
limiting rather than overloading of the amplifier – this scheme
yields a cleaner sinusoidal output signal.
The phase criterion in is met using a resonant circuit in
the feedback network. Why? Because near the resonant
frequency of the feedback circuit, the phase ?L varies rapidly,
as shown in This characteristic allows precise
placement of the oscillator frequency. Clever! (Also has the
effect of producing smaller “phase noise.”)
Hence, from the two curves in we see that the
oscillation criteria are met when
• G( P0) = L (at a certain P0)
• ?L( f0) = ?G (at a certain f0) Oscillator Startup
Another important aspect of oscillators is how they begin
oscillating (remember: no input!). The criteria we just derived
apply to steady state power at the frequency of oscillation.
There are two general approaches to starting an oscillator: (1)
repeated amplification of noise, or (2) with an external startup
signal (as in super-regenerative receivers).
If|G| > |L |, then noise that meets the phase criterion (11.6) will
be repeatedly amplified. At startup, we will use the small signal
gain g to state the start-up criterion for feedback oscillators:
1. |g| >|L| ,
2. ?L( f0) = ?g .
Interestingly, some oscillators that work well at relatively high
power will not start-up by themselves at low power. An example
of this is Class C amplifiers, like the Power Amplifier Q7 in the
NorCal 40A. Rather than the gain curve shown in Fig. 11.1(b), class C amplifiers have the gain curve shown below. (This G curve was constructed from data collected in Prob. 24.B.)
Class C amplifiers will not oscillate if P < Ps . However, once P > Ps oscillation may occur if the feedback network meets the phase criterion It turns out, interestingly, that the oscillators in the NorCal 40A (such as the VFO) actually startup in Class A then shift to Class C as P increases. Clapp Oscillator
There are many topologies for feedback oscillator circuits.
However, all can be divided into two general classes: (1)
Colpitts and (2) Hartley oscillators. Each contains an amplifier,
a resonator and a voltage divider network to feed some of the
output signal back to the input (called “feedback”).
In Colpitts oscillators, capacitors form the voltage divider, while
inductors form the divider network in Hartley oscillators
The VFO in the NorCal 40A is a Clapp oscillator, which is a
member of the Colpitts family since capacitors form the voltage
divider network:
We will analyze the NorCal 40A VFO in two stages. First is
startup using small-signal (i.e., linear) analysis. In the next
lecture, we will look at steady state using a large signal analysis VFO Startup Condition
We can construct the small signal equivalent circuit for the VFO
in as shown in
Referring back to the input x in this circuit is vgs while the output y is id :
With id = gm vgs for the JFET, we use the small-signal gain gm = g in the startup criterion and
1. |gm| >| L| ,
2. ?L f0 = ?g0 .
gmm is a real and positive quantity dependent on the type of JFET and the value of vgs. See (p. 173) for an example We’ll now solve for vgs in terms of id, since L is a ratio of them. This circuit will oscillate at the resonant f of the tank because of the phase criterion (11.9). The resonant frequency f0 is
.
which is in series with L1. Now, at this resonant f = f0, i2 + i = 0 (a key!). Hence, with
Therefore,
At resonance, the source terminal of the JFET has the voltage
vs = Rid
Substituting gives
This is our needed equation since we have vgs in terms of id . Now, by the definition of L in
By obtaining this equation, we have solved for the small signal
loss factor of the feedback network in Fig. 11.5.
With this L factor now known, it is simple matter to determine
the startup condition for the VFO. Specifically, we find that the startup condition for this JFET VFO (Clapp oscillator) is
In the NorCal 40A, C1 = C2 (actually C52 = C53) giving the startup condition
But what about the phase condition ?L( f0) = ?gm ? Notice that both gm and L have zero phase shift at the resonant frequency. Consequently, the phase criterion for startup is intrinsically satisfied. In summary, if the condition (11.25) is satisfied the VFO circuit
in the NorCal 40A will begin to oscillate on its own by repeatedly amplifying noise. Very cool! Check VFO Startup Design
Let’s carefully look at VFO startup in the NorCal 40A. The load
resistance R of is R23. The VFO begins oscillation
with vg near zero because of R21, which is why it is called the
“start up resistor.” With vgs £ 0, then gm is large:
Now let’s check the startup condition. From
At startup, 18 gm ? mS while 1/R23 ? 0.556 mS. The answer is then yes (by 32x). Therefore, the VFO in the NorCal 40A should easily start up.
.
